On Monday 05 October 2009, afra at aurigae.demon.co.uk wrote: NCR also made Microchannel machines, including some quite large ones (Worldmark 5100, which at some point I had a couple of). Pat -- Purdue University Research Computing --- http://www.rcac.purdue.edu/ The Computer Refuge --- http://computer-refuge.org

On Mon, Oct 5, 2009 at 10:26 AM, Patrick Finnegan <pat at computer-refuge.org> wrote: I worked with some large NCR MCA machines in the late 1990s when I was at Lucent. Imagine a machine larger than a VAX 11/750 with four Pentium-90 processors, multiple MCA buses and many SCSI cards on those MCA buses and many, many disks... the scary part was it was still so PC-compatible that it could still boot DOS and run Doom. -ethan

Quoting "Michael B. Brutman" <mbbrutman-cctalk at brutman.com>: Is that Dos 7 as in Win 9x dos ? or as in PC-DOS 7 ? If the former, odd that it doesn't have any 386 instructions in there as 9x of course required a 386 or above. Guess if you used OpenDos (or Win98 dos 7), you can use Fat32 and have Phill. ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

On Oct 5, 2009, at 12:04 PM, Chuck Guzis wrote: Yes, unfortunately this is true. It's highly annoying for those of us who like to interface IDE drives to 8-bit processors but don't want to use CF. Ahh, very nice. Sorta like the very nice GIDE interface for Z80 systems. I've designed a similar circuit in Verilog, but haven't tested it in real hardware yet. Yes, they do. CF cards are required to implement the 8-bit transfer mode. -Dave -- Dave McGuire Port Charlotte, FL

----- Original Message ----- From: "Steven Hirsch" <snhirsch at gmail.com> To: "General Discussion: On-Topic Posts Only" <cctech at classiccmp.org> Sent: Monday, October 05, 2009 6:18 PM Subject: Re: XTIDE controller now available I would think an 8bit SCSI version would be cool to have as well, still plenty of smaller SCSI drives around as well as Iomega and Syquest removables plus tape drives.

On Oct 6, 2009, at 3:24 PM, Fred Cisin wrote: Quite. 16-bit ISA boards are all over the place though, like the venerable Adaptec 1542 series. I wonder if they will work in an "overhang" configuration in 8-bit slots; I've never tried. -Dave -- Dave McGuire Port Charlotte, FL

----- Original Message ----- From: "Fred Cisin" <cisin at xenosoft.com> To: "General Discussion: On-Topic and Off-Topic Posts" <cctalk at classiccmp.org> Sent: Tuesday, October 06, 2009 3:24 PM Subject: Re: XTIDE controller now available Yes, especially ones with a bootable ROM that would work on an XT. I have a few 8 bit ISA SCSI cards without ROM meant to connect CDROMs and other devices like scanners (using drivers).

On Oct 5, 2009, at 3:39 PM, Geoffrey Reed wrote: The big difference here is LBA vs. C/H/S addressing. Recall that the IDE interface is a clone of the register-level programming interface of the ancient WD1010 MFM hard disk controller chip family, whose registers are 8 bits wide. There's a register for the starting sector number (256 sectors), a register for the low byte of the cylinder address and two bits of another register for the high part of the cylinder address (meaning it tops out at 1024 cylinders), and half of a register for the head number (meaning 16 heads). 1024 cylinders * 16 heads * 256 sectors * 512 bytes/sector = 2GB I don't know where any supposed 520MB limit comes from; if someone can enlighten me there, I'd appreciate it. LBA (Logical Block Addressing) mode takes a large, flat linear address space of blocks with no notion of cylinders, heads, or sectors. It's 28 bits wide, topping out at about 137GB. IDE interfaces take that word of bits, break it up into little pieces, and sprinkle them across the former cylinder/head/sector registers (and pieces of registers) in a very kludgy and inefficient way. They take the LBA address and put the first 8 bits into the sector register, the middle 16 bits into the cylinder number register (yes, using it as a 16-bit register!), and the high four bits go into the head and device select register. Ugly, ugly, ugly. New kludges were devised to exceed this ~137GB plateau (obviously), but I don't know how they work offhand because I've never designed with them. Anyway, early BIOSs don't know how to deal with this, which is one of the things that makes this XTIDE controller special. -Dave -- Dave McGuire Port Charlotte, FL

Dave McGuire wrote: IIRC it's from the fact that most BIOSes can only deal with a 6 bit sector number thus : AH = 02h AL = number of sectors to read (must be nonzero) CH = low eight bits of cylinder number CL = sector number 1-63 (bits 0-5) high two bits of cylinder (bits 6-7, hard disk only) DH = head number DL = drive number (bit 7 set for hard disk) ES:BX -> data buffer Copied from Ralf Brown's Interrupt List So 1024 cylinders x 16 Heads x 64 sectors is 512MB Later bioses I believe used the top bits of the head number to get greater capacity. Cheers. Phill. -- Phill Harvey-Smith, Programmer, Hardware hacker, and general eccentric ! "You can twist perceptions, but reality won't budge" -- Rush.

On Oct 5, 2009, at 2:10 PM, Dave McGuire wrote: It was originally 63 sectors. I was one of the 3 people at IBM who came up with this scheme (because the controller we had did LBA addressing so we needed a way to map C/H/S to LBA in a simple way). We *knew* at the time that this was constraining but we figured that LBA would be in place before we hit the limit. We were wrong. TTFN - Guy